In certain cellular networks, there may be a large disparity between the average cell data throughput. For example, it has been observed that certain Long Term Evolution (LTE) networks experience a 10:1 ratio between cell-average throughput and cell-edge throughput. Such a disparity may result in different users experiencing significant differences in the difference in Quality of Service (QoS).
Embodiments of the invention described include a feedback mechanism for a wireless network (e.g., an LTE cellular network) in which a wireless device (e.g., an LTE UE) transmits downlink interference data to a first transmission point (TP) (e.g., an LTE eNB that serves the UE). The wireless device generates this information by listening to reference signals (e.g., channel state information reference signals (CSI-RS)) received from the first TP as well as from a second TP (e.g., an eNB that is interfering with the signals from the serving eNB). The wireless device uses these reference signals to estimate a first channel matrix relating to the first TP (e.g., a channel response matrix H11 relating to the signals passing between the UE antennas and the eNB antennas) and a second channel matrix relating to the second TP. Using the estimated first and second channel matrices, the wireless device derives pairs of precoding matrices. The wireless device attempts to choose a rank and a particular pair of precoding matrices so as to maximize the mutual information or sum rate (e.g., the sum of mutual information across different transmission layers or streams). Once it chooses the appropriate pair of precoding matrices for the appropriate rank, the wireless device transmits this information to the first TP (e.g., transmits the information in the Rank Indicator (RI) and the Precoding Matrix Indicator (PMI) along with the Channel Quality Indicator (CQI) in the Channel State Information (CSI) feedback report). The network may use this information to adjust the transmissions of various TPs (e.g., an LTE network may adjust beam direction or precoding at one or more eNBs so as to maximize total network throughput or some other network utility such as proportionally fair (PF) metric, max-min rate metric, etc.). To gather the information that enables this adjustment, the network may collect data from multiple TPs in the network. There may be a centralized controller for a subset of TPs within the network (i.e., the TPs that comprise the coordination area) such coordinated transmission is possible. Coordinated transmission includes coordinated resource allocation, coordinated precoding and spatial multiplexing (including joint transmission (JT), dynamic point selection (DPS), dynamic point blanking (DPB), coordinated beamforming (CB), etc.), coordinated power control, where each of these decisions are taken in a coordinated manner with the exchange of CSI between the individual TPs in the coordination area (i.e., cells under the coordinating TPs controlled the centralized controller) and the centralized controller.
In other embodiments of the invention, a first TP (e.g., an LTE eNB) configures a wireless terminal (e.g. an LTE UE) to report channel state information feedback. The first TP receives, from the UE, information relating to a pair of precoding matrices, wherein a first precoding matrix of the pair is applicable to a second TP and represents an assumption on the desired signal transmitted from the second TP. A second precoding matrix of the pair is applicable to a third TP and represents an assumption on the interference signal from the third TP. The network controller (which may reside in the first TP, the second TP or elsewhere within the network) determines a precoding weight matrix based on the first precoding matrix and the second precoding matrix for transmission of information from the second TP. The second TP transmits encoded information bits to the wireless terminal from its antenna ports based on the precoding weight matrix.
Although the various embodiments are frequently described in the context of an LTE cellular system, it is to be understood that the scope of the invention is not limited to LTE and may be implemented in other types of wireless networks (IEEE 802.11, 802.16, etc.).
Before going into further detail regarding various embodiments of the invention, some general concepts will now be discussed. It is to be understood that the discussion of these concept is not intended to limit the scope of the invention, but rather to aid in describing certain embodiments.
The concept of “beamforming” according to an embodiment of the invention will now be introduced. Beamforming is a general signal processing technique used to control the directionality of the reception or transmission of a signal on an array of transmitters or receivers. Beamforming takes advantage of the following properties: The transmitters or receivers in the array may be physically located on a single device or may be distributed on multiple devices. Using beamforming, a device can direct the majority of its signal energy from a group of transmitters (such as radio antennas) in a particular angular direction. Similarly, a device can use beamforming so that its receivers signal from a particular angular direction. When multiple transmitters are located near one another and send out signals, an antenna gain pattern may occur where the signals combine destructively. If, however, the delay and/or phase in the signals of the various transmitters are chosen in an appropriate manner, a beneficial antenna gain pattern may be created where the signals from the various transmitters combine constructively at least at one angular direction. The receive case works the same way, except that signals are being received rather than transmitted.
The concept of “precoding” according to an embodiment of the invention will now be introduced. Precoding is based on transmit beamforming concepts with the provision of allowing multiple beams to be simultaneously transmitted in a Multiple-Input Multiple-Output (MIMO) system. For example, the LTE specification defines a set of complex-valued weighting matrices for combining the transmission layers before transmission using various antenna configurations. The set of all weighting matrices is referred to as a “codebook,” and generally each element (weighting matrix) of the codebook is a precoding matrix or a precoding vector. Codebooks may be “finite” in the sense that there are a finite number of elements from which to choose. For example, codebooks in LTE have a finite number of elements (e.g., a 4 transmit antenna codebook has 16 elements where each element can be represented by 4 bits).
The concept of a “precoding matrix” according to an embodiment of the invention will now be introduced. If a transmitter (e.g., a TP) is to send S streams over N antennas to a receiver (e.g., one or more UEs), the transmitter can use a linear transformation P on s, to obtain a N×1 signal vector x in the form:x=Ps, wherein P is a N×S complex-valued matrix and s is the S×1 signal vector.
Each element of s corresponds to one of the S streams that the transmitter intends to transmit. An element of P is P[n,s] where P[n,s] is the element in the n-th row and s-th column of P that describes the weighting coefficient associated with the s-th stream on the n-th antenna. For example, if S=1 and N=2, P[1,1] corresponds to the weighting coefficient (or the beamforming factor) associated with the (single) stream that is being transmitted on the first antenna and P[2,1] corresponds to the weighting coefficient associated with the stream on the second antenna. P[1,1] and P[2,1] can be chosen by the transmitter so as to maximize the signal power transmitted in a certain direction of the receiver. The linear transformation P is referred to as a “precoder” or a “precoding matrix.” In wireless systems that make use of transmission based on closed-loop feedback, P belongs to a finite set of precoding matrices C, also known as the codebook. In other words, PεC in set-theoretic notation.
The concept of a dual-codebook structure for precoding according to an embodiment of the invention will now be introduced. In this structure, P=W1W2 where W1εC(1) and C(1) is a first codebook; and W2εC(2) where C(1) is a second codebook. The first precoding matrix W1 attempts to capture the wideband spatial properties (over all subcarriers within all subbands of interest) of the channel and the second precoding matrix W2 attempts to capture the subband spatial properties (relative to the chosen W1) of the channel. The roles of W1 and W2 can be reversed (i.e. W1 represents subband spatial properties and W2 represents wideband spatial properties).
The concept of a “channel response matrix” according to an embodiment of the invention will now be introduced. The signal transmitted from the N antennas by a transmitter propagates through a radio channel and are received by M antennas at the receiver. Under some conditions (e.g., the channel does not have memory or there is inter-symbol interference and the channel is flat-fading), the received signal vector y of size M×1 can be written as a linear transformation of the transmitted vector:y=Hx, 
where H is a complex-valued M×N matrix also known as the channel response matrix. Each element of H is H[m,n], which is the element in the m-th row and n-th column of H, describes the channel complex gain associated with the radio link between the n-th transmit antenna and m-th receive antenna. For example, if there are M=2 receive antennas and N=1 transmit antenna, H[1,1] corresponds to the channel gain between the transmit antenna and the first receive antenna and H[2,1] corresponds to the channel gain between the transmit antenna and the second receive antenna.
The concept of an antenna port according to an embodiment of the invention will now be introduced. An “antenna port” may be a logical port that may correspond to a beam (resulting from beamforming) or may correspond to a physical antenna at a User Equipment (UE) or a Transmission Point (TP). An antenna port may be defined such that a channel over which a symbol on the antenna port is conveyed can be inferred from the effective channel over which another symbol on the same antenna port is conveyed. More generally, an antenna port can correspond to any well-defined description of a transmission from one or more of antennas. For example, it could include a beamformed transmission from a set of antennas with appropriate antenna weights being applied, where the set of antennas itself could be unknown to a UE. In some particular implementations “antenna port” can also refer to a physical antenna port at the TP. In certain cases, the beamforming or precoding applied at the TP may be transparent to the UE. In other words, the UE need not know what precoding weights are used by the TP for a particular transmission on the downlink. In this disclosure, the term “antenna” is used as a shorthand for “antenna port.” Thus an “antenna” as referred to herein may, in fact, be a virtual antenna that includes multiple antennas that function as a single antenna. It may also be a single antenna.
The concept of a “layer” according to an embodiment of the invention will now be introduced. Each logical path between an antenna port and a UE will be referred as a “layer.” The number of layers between an antenna port and a UE is referred to as its “Rank.” Thus, a Rank 1 transmission has one layer, Rank 2 has 2 layers, and so on.
Some embodiments of the invention are described herein using matrix math concepts, which will now be briefly described. The singular value decomposition (SVD) of a matrix is a factorization of a real valued or complex-valued matrix. Formally, the singular value decomposition of an m×n real or complex matrix A is a factorization of the formA=UΣV* 
where U is an m×m real or complex unitary matrix, Σ is an m×n rectangular diagonal matrix with nonnegative real numbers on the diagonal, and V* (the conjugate transpose of V) is an n×n real or complex unitary matrix. The diagonal entries Σi,i of Σ are known as the singular values of A. The m columns of U and the n columns of V are called the left-singular vectors and right-singular vectors of A, respectively.
A column space may be described as follows. Let A be a complex-valued m×n matrix, with column vectors are v1, v2, . . . , vn. A linear combination of these vectors is any vector of the formc1v1+c2v2+ . . . +Cnvn where c1, c2, . . . , cn are complex scalars. The set of all possible linear combinations of v1, . . . , vn is called the column space of A denoted as col(A). That is, col(A) is the span of the vectors v1, . . . , vn.The nullspace (or kernel) of an m×n matrix A is the setN(A)=Null(A)=Ker(A)={xεCn:Ax=0},
where 0 denotes the zero vector with m components. The matrix equation Ax=0 is equivalent to a homogeneous system of linear equations:
  Ax  =      0    ↔                                                                      a                11                            ⁢                              x                1                                      +                                                                              a                12                            ⁢                              x                2                                      +            …            +                                                                              a                                  1                  ⁢                  n                                            ⁢                              x                n                                      =                                    0                                                                                a                21                            ⁢                              x                1                                      +                                                                              a                22                            ⁢                              x                2                                      +            …            +                                                                              a                                  2                  ⁢                  n                                            ⁢                              x                n                                      =                                    0                                      ⋮                          ⋮                          ⋮                          ⋮                                                                                a                                  m                  ⁢                                                                          ⁢                  1                                            ⁢                              x                1                                      +                                                                              a                                  m                  ⁢                                                                          ⁢                  2                                            ⁢                              x                2                                      +            …            +                                                                              a                mn                            ⁢                              x                n                                      =                                    0.                    
Turning to the drawings, an example of a network in which an embodiment of the invention operates will now be described. FIG. 1 illustrates a communication system 100 including a network 102, TP1, TP2 and TP3 (which may be implemented as part of Enhanced Nodes B (eNBs) or of Remote Radio Heads (RRHs)), and user equipment UE1, UE2 and UE3. Various communication devices may exchange data or information through the network 102. The network 102 may be an evolved universal terrestrial radio access (E-UTRA) or other type of telecommunication network. The network 102 may be capable of using Coordinated Multipoint (CoMP) techniques. CoMP will be described in more detail below.
The TPs may be linked to one another by a series of land and wireless connections that will be generally referred to as a “backhaul network.” In an embodiment, a TP may be a distributed set of servers in the network 102. In another embodiment, a TP may correspond to a set of geographically collocated or proximal physical antenna elements. Each TP provides communication services to UEs in a geographic/site area which may be sectorized in to one or regions referred to as a cell or intra-site cell. For example, a TP may serve three cells, in which each cell can be considered as having a TP.
A UE may be one of several types of devices, such as, a mobile phone (e.g., smartphone), or laptop. Each TP may have one or more transmit antennas and one or more receive antennas. Similarly, a UE may have multiple receive antennas communicating with one or more of the TPs. Each antenna port may transmit or receive signals. A multiple antenna configuration on either a TP or a UE is may support Multiple Input Multiple Output (MIMO) communication.
To send data to a UE, a TP precodes the data to be communicated onto the one or more layers, maps and transmits the resulting precoded data to one or more antenna ports. The effective channel (or beamformed channel) corresponding to a transmission layer may be estimated based on reference signals mapped to one or more antenna ports.
FIG. 2 illustrates a possible configuration of a computing system to act as a TP (e.g., one or more of the TPs in FIG. 1). The TP may include a processor/controller 210, a memory 220, a database interface 230, a transceiver 240, input/output (I/O) device interface 250, and a network interface 260, connected through bus 270. The TP may implement any operating system, such as Microsoft Windows®, UNIX, or LINUX, for example. Client and server software may be written in any programming language, such as C, C++, Java or Visual Basic, for example. The server software may run on an application framework, such as, for example, a Java® server or .NET® framework.
The processor/processor 210 may be any programmable processor. The subject of the disclosure may also be implemented on a general-purpose or a special purpose computer, a programmed microprocessor or microprocessor, peripheral integrated circuit elements, an application-specific integrated circuit or other integrated circuits, hardware/electronic logic circuits, such as a discrete element circuit, a programmable logic device, such as a programmable logic array, field programmable gate-array, or the like. In general, any device or devices capable of implementing the decision support method as described herein may be used to implement the decision support system functions of this disclosure.
The memory 220 may include volatile and nonvolatile data storage, including one or more electrical, magnetic or optical memories such as a random access memory (RAM), cache, hard drive, or other memory device. The memory may have a cache to speed access to specific data. The memory 220 may also be connected to a compact disc-read only memory (CD-ROM), digital video disc-read only memory (DVD-ROM), DVD read write input, tape drive, or other removable memory device that allows media content to be directly uploaded into the system. Data may be stored in the memory 220 or in a separate database. The database interface 230 may be used by the processor/controller 210 to access the database. The database may contain any formatting data to connect UE to the network 102 (FIG. 1). The transceiver 240 may create a data connection with the UE.
The I/O device interface 250 may be connected to one or more input devices that may include a keyboard, mouse, pen-operated touch screen or monitor, voice-recognition device, or any other device that accepts input. The I/O device interface 250 may also be connected to one or more output devices, such as a monitor, printer, disk drive, speakers, or any other device provided to output data. The I/O device interface 250 may receive a data task or connection criteria from a network administrator.
The network connection interface 260 may be connected to a communication device, modem, network interface card, a transceiver, or any other device capable of transmitting and receiving signals from the network 102. The network connection interface 260 may be used to connect a client device to a network. The network connection interface 260 may be used to connect the teleconference device to the network connecting the user to other users in the teleconference. The components of the TP may be connected via an electrical bus 270, for example, or linked wirelessly.
Client software and databases may be accessed by the processor/processor 210 from memory 220, and may include, for example, database applications, word processing applications, as well as components that embody the decision support functionality of the present disclosure. A TP (FIG. 1) may implement any operating system, such as Microsoft Windows®, LINUX, or UNIX, for example. Client and server software may be written in any programming language, such as C, C++, Java or Visual Basic, for example. Although not required, the disclosure is described, at least in part, in the general context of computer-executable instructions, such as program modules, being executed by the electronic device, such as a general purpose computer. Generally, program modules include routine programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Moreover, those skilled in the art will appreciate that other embodiments of the disclosure may be practiced in network computing environments with many types of computer system configurations, including personal computers, hand-held devices, multi-processor systems, microprocessor-based or programmable consumer electronics, network PCs, minicomputers, mainframe computers, and the like.
FIG. 3 illustrates in a block diagram one embodiment of a telecommunication apparatus or electronic device to act as a UE (such as one or more of the UEs depicted in FIG. 1). The UE may be capable of accessing the information or data stored in the network 102. For some embodiments of the disclosure, the UE may also support one or more applications for performing various communications with the network 102.
The UE may include a transceiver 302, which is capable of sending and receiving data over the network 102. The UE may include a processor 304 that executes stored programs. The UE may also include a volatile memory 306 and a non-volatile memory 308 which are used by the processor 304. The UE may include a user input interface 310 that may comprise elements such as a keypad, display, touch screen, and the like. The UE may also include a user output device that may comprise a display screen and an audio interface 312 that may comprise elements such as a microphone, earphone, and speaker. The UE also may include a component interface 314 to which additional elements may be attached, for example, a universal serial bus (USB) interface. Finally, the UE may include a power supply 316.
Referring again to FIG. 1, the general mode of communication between a TP and a UE according to an embodiment of the invention will now be described. A TP and a UE generally communicate via uplink channels and downlink channels. The physical medium used is Radio Frequency (RF) signals, which are encoded using Orthogonal Frequency-Division Multiplexing (OFDM). The modulation scheme used by the TP and the UE differs depending on whether the signals are being sent in the uplink direction (which is the UE to TP direction) or the downlink direction (which is the TP to UE direction). The modulation method used in the downlink direction is a multiple-access version of OFDM called Orthogonal Frequency-Division Multiple Access (OFDMA). In the uplink direction, Single Carrier Frequency Division Multiple Access (SC-FDMA) is used.
The data carried by the OFDM signals in the downlink direction is organized into radio frames. Each radio frame typically includes ten subframes. FIG. 4 depicts a time-frequency diagram of a downlink subframe 400. It is to be understood that the subframe 400 shows one example subframe format, and that other formats are possible.
The vertical scale of FIG. 4 represents frequency, while the horizontal scale represents time. The vertical scale is divided into multiple blocks of frequency, or OFDM subcarriers (“subcarriers”) that may be allocated for transmission. The horizontal scale of FIG. 4 is divided into multiple blocks of time, or OFDM symbols (“symbols”) that may be allocated for transmission. The subframe 400 is divided into time-frequency resource blocks (RBs) (e.g., Resource Block 0 (RB0), Resource Block 1 (RB1), etc.). Each RB is twelve subcarriers by seven symbols. The subframe 400 is a total of 1.0 ms long and is divided into two time slots of 0.5 ms each. In turn, each RB can be divided into multiple resource elements (REs). Each RE is a one subcarrier by one symbol. It is to be noted that many downlink subframes may be transmitted from the TP to the UE and that various channels may occupy slots in many sub-frames.
The subframe 400 may include several types of reference signals in an embodiment of the invention. The references signals are transmitted by the TP to the UE to enable the UE to perform various functions. One such reference signals, which may be used in an LTE embodiment of the invention, will now be introduced. A Channel State Information Reference Signal (CSI-RS) is used by the UE to determine channel-state information (CSI) that the UE reports to the TP. The CSI-RS is not necessarily transmitted all subframes.
Referring again to FIG. 4, other reference signals that the subframe 400 may contain in an embodiment of the invention will now be described. REs labeled R7-R10 (and associated with antenna ports 7-10, respectively) are allocated to a Demodulation Reference Signal (DMRS) with the REs being referred to as DMRS REs. Typically, reference signals corresponding to antenna ports 7 and 8 are multiplexed using Code Division Multiplexing (CDM) or other scheme and are mapped to the same REs in time and frequency domain. The subframe can also include other reference signals such as cell-specific reference signal (CRS), positioning reference signal (PRS), primary synchronization signal (PSS) and secondary synchronization signal (SSS) that are distributed in the control regions and/or user data regions of the sub-frame. These other reference signals may be present but are not necessarily used for demodulation of received signals by a UE in an LTE-A communication system. For example, the other reference signal may include the CSI-RS, muted (or zero-power) CSI-RS where the UE can assume and zero transmission power on the reference signal REs that may be useful for interference measurements, improving channel measurements on CSI-RS from other TPs etc. The CSI-RS is typically not used for demodulation purposes and may be present in occasional sub-frames, i.e., the sub-frame periodicity, sub-frame offset (relative to a radio frame boundary), and the number of CSI-RS antenna ports are configurable via higher layer signaling. CSI-RS typically occupy REs that are not occupied by CRS, potential DMRS, etc.
The CSI that the UE reports to the TP includes a Channel Quality Indicator (CQI), a Pre-coding Matrix Indicator (PMI), a Rank Indicator (RI), and a Precoding Type Indicator (PTI). The CQI tells the TP information about the link adaptation parameters that the UE can support at that time, taking into account the transmission mode, the receiver type of the UE, the number of antennas being used by the UE and the interference being experienced by the UE. The UE reports back to the TP the highest CQI index corresponding to the MCS and TBS (Transport Block Size) for which the estimated received download transport block Block Error Rate (BLER) (e.g., the ratio of blocks sent with transmission errors to total blocks sent) does not exceed 10%. The PMI indicates to the TP the precoding matrix that the UE recommends the TP use for data transmission over multiple antennas. The RI is the recommendation of the UE to the TP for the number of transmission layers to be used in spatial multiplexing. The RI can have, for example, values of 1 or 2 with a 2-by-2 antenna configuration and from 1 to 4 with a 4-by-4 antenna configuration. The PTI distinguishes slow from fast fading environments. When the UE is configured to report PMI based on a dual-codebook, the PTI indicates whether the reported PMI corresponds to a first precoding matrix type (e.g., wideband) or a second precoding matrix type (e.g., subband).
According to an embodiment of the invention, the CSI-RS may occupy a set of time-frequency resources (e.g., a particular “pattern” of REs) of a subframe. This set of time-frequency resources will be referred to collectively as a “CSI-RS resource.” A CSI-RS resource may be described by a “CSI-RS configuration.” The CSI-RS configuration may include one or more of the following:
A subframe configuration, such as the identity of the subframes on which the CSI-RS is transmitted, a subframe offset (relative to a known reference point in time), and the periodicity of the subframes in which CSI-RS are transmitted;
A resource configuration, such as the OFDM symbols and corresponding REs of the subframes in which CSI-RS is transmitted; and
The identity and number of antenna ports being used to transmit the CSI-RS (e.g., # of ports=4, identity of ports=15, 16, 17, and 18).
In an embodiment of the invention, the TP provides the CSI-RS configuration to the UE via RRC signaling. The RRC layer in the UE provides the CSI-RS configuration information to the physical layer in the UE (e.g., “higher layer signaling”).
It is to be noted that there may be a one-to-one correspondence between each TP and each CSI-RS resource, or each CSI-RS resource may represent transmission from multiple TPs (such as when two or more TPs are involved in Joint Transmission (JT) or Dynamic Point selection (DPS)). Additionally, each CSI-RS-resource may have one or more antenna port associated with it. These possibilities hold true for the CSI-RS. There may be a one-to-one correspondence between CSI-RS and TPs, one-to-many, or one CSI-RS per antenna port.
The following is an example of a data structure that may be used for to express a CSI-RS configuration:
CSI-RS-ConfigThe IE CSI-RS-Config is used to specify the CSI (Channel-StateInformation) reference signal configuration.CSI-RS-Config information elementsASN1STARTCSI-RS-Config-r10 ::= SEQUENCE { csi-RS-r10    CHOICE {  release NULL,  setupSEQUENCE {  antennaPortsCount-r10  ENUMERATED {an1, an2, an4, an8},  resourceConfig-r10INTEGER (0..31),  subframeConfig-r10INTEGER (0..154),  p-C-r10    INTEGER (−8..15)  } }OPTIONAL,-- Need ON zeroTxPowerCSI-RS-r10 CHOICE {  releaseNULL,  setupSEQUENCE {  zeroTxPowerResourceConfigList-r10 BIT STRING (SIZE (16)),  zeroTxPowerSubframeConfig-r10 INTEGER (0..154)  } }OPTIONAL-- Need ON}ASN1STOP
Although the UE receives the identity of and the number of the antenna ports by which the CSI-RS is transmitted, the UE does not know which antenna port is assigned to which TP. The antenna port to TP mapping is maintained by the network.
To enable communication to occur smoothly, the TP (FIG. 1) uses control signaling, including downlink (DL) signaling via DL control channel, and uplink (UL) signaling, via UL control channel. A DL control channel may include Downlink Control Information (DCI), which may include a downlink assignment, an uplink grant, or an uplink power control command. A downlink assignment may include one of more of a downlink resource allocation information, DL Hybrid Acknowledge Repeat Request (HARQ) information, DL MIMO information, power control information, user identifier or Radio Network Temporary Indentifier (RNTI), etc. Similarly, the UL grant may include uplink resource allocation information, uplink HARQ information, uplink MIMO information, power control commands, user identifier or RNTI, etc. Some types of uplink grants are used for causing a UE to report specific control information such as channel quality information (CQI) or channel state information (CSI). Other control information that the UE may send in response to this type of uplink grant includes UE identify information, buffer occupancy information, and power control state information. A TP may request CSI information (e.g., using an uplink grant) periodically or aperiodically.
An embodiment of the invention may be carried out in a wireless network with multiple transmission points (TPs), in which two or more of the TPs coordinate to act as a single transmission point. One such coordination scheme that may be used in an LTE implementation is Coordinated Multi-Point (CoMP). CoMP allows semi-static or dynamic coordination or transmission and reception with multiple geographically separated TPs.
Four possible CoMP deployment scenarios will now be described with reference to FIG. 5. The first scenario is a homogeneous cellular network with intra-site CoMP. A central entity (such as the TP or an entity within the TP such as a scheduler) can control or coordinate the transmissions from the TPs of each of the three cells within its geographic area or coordination area which is referred to as intra-site CoMP. The second scenario is a homogeneous network with high transmission power Remote Radio Heads (RRHs). A RRH may be passive amplifiers, or each may contain full signal processing capability (i.e., a transceiver). RRH can also be referred to as a sub-base station, remote antenna unit, Remote Radio Unit (RRU) or other terminology known in the art. The geographic coverage area of a RRH may be sectorized in to one or cells, with each cell can be considered as having a TP. In the second scenario, the central entity can coordinate nine cells as a baseline. The designer of a system according to the second scenario may choose between 3, 19, 21 cells as a potential optional value.
The third scenario, is a heterogeneous network with low power RRHs (or low power nodes (LPN)) within the macrocell coverage. The macrocell may be an eNB intra-site cell and may be considered as being associated with a TP. The geographic coverage area of each low power RRH in the third scenario may be generally referred to as a subcell, and the RRH (with no sectorization) can be considered as being a TP. The RRHs are geographically (or spatially) separated and controlled by the TP of the macrocell. The TP can be considered as using a Distributed Antenna system (DAS) with the macrocell TP and each RRH associated with one or more antenna ports in the DAS. In this scenario, transmission/reception points created by the RRHs have different cell IDs as the macro cell. The coordination area includes 1 cell with N low-power nodes as starting point, and can include 3 intra-site cells with 3*N low-power nodes.
The fourth scenario, which is depicted in the same drawing as the third scenario, in a cellular network having low power RRHs within the macrocell coverage, similar to third scenario except the transmission/reception points created by the RRHs have the same cell IDs as the macro cell. The coordination area includes one cell with N low-power nodes as starting point and can include three intra-site cells with 3*N low-power nodes.
Although many CoMP scenarios are possible, some of the embodiments described herein are particularly directed to the second, third, and fourth CoMP deployment scenarios. The invention may be implemented in these scenarios, with the eNBs (or RRHs) that are depicted acting as TP1, TP2, etc. and one or more UEs (not shown) operating within the coverage area of one or more of the eNBs (or RRHs).
When implemented using 4G LTE, CoMP may employ a transmission scheme known as Joint Processing/Joint Transmission (JP/JT). In JP/JT, data for a UE is available at more than one point and there is coordination between multiple TPs that are geographically separated and that are simultaneously transmitting to or receiving from UEs. One type of JP/JT is Dynamic Point Selection/Dynamic Point Blanking (DPS/DPB). In DPS/DPB, certain of the transmitters are selected to transmit, while others are selected to refrain from transmitting (i.e., are selected for blanking). The transmitting/blanking point may change from one sub-frame to another including varying over the RB pairs within a sub-frame. Data is available simultaneously at multiple points. DPS may be combined with JT in which case multiple points can be selected for data transmission in the time-frequency resource.
Referring again to FIGS. 1 & 4, TPs and UEs may communicate with one another under a variety of channel conditions. The embodiments described herein are not limited to any particular set of channel conditions.
In order to illustrate the underlying concepts more clearly, it will be assumed that each channel is varying slowly enough so that the signal subspace which is col(H11) and interference subspace which is col(H12) predominantly remain the same from one transmission (e.g., subframe n) to the next transmission (subframe n+8), so as to allow, for example, the UE1 to send feedback to TP1 and TP2 so that TP1 and TP2 can choose appropriate precoders in a coordinated manner.
Referring again to FIG. 1, the received signal at UE1 can be written asy1=H11P1x1+H21P2x2+n1  (1)and the received signal at UE2 can be written asy2=H22P2x2+H12P1x1+n2,  (2)where
Hi,j is the channel response matrix when TP i transmits and UE j receives;
Pi is the precoding matrix (or vector) employed by TP I;
nj˜σjCN(0, I) is the residual co-channel interference and noise vector at UE j.
σj2 is the variance associated with the residual co-channel interference on the j-th UE.
In some embodiments of the invention, each UE employs a linear Minimum Mean Square Estimation (MMSE) receiver to recover the transmitted signal vector. For UE1, for example:{circumflex over (x)}1=G1y1, where G1=P1*H11*(σ12I+H21P2P2*H21*+H11P1P1*H11*)−1  (3)where Gn is the receive filter (spatial) applied to the signal vector received at the n-th UE. There may be a similar expression for UE2.
If UE1 uses a linear MMSE filter, the filter can be equivalently written asG1=(I+P1*H11*R1−1H11P1)−1P1*H11*R1−1, where  (4)R1=σ12I+H21P2P2*H21*.  (5)
If one or more of the UEs is experiencing rank-1 effective interference (i.e., rank of the effective channel matrix of the interference component H21P2 is 1), using the matrix inversion lemma, the following result occursR1−1=σ1−2(I−h21h21*/(σ12+|h21|2)) where h21 is a vector such that  (6)H21P2P2*H21*=h21h21*.  (7)
When the residual co-channel interference becomes negligible (σ1→0), the following is obtained:G1=(P1*H11*Ph21⊥H11P1)−1P1*H11*Ph21⊥  (8)where Ph21⊥ is the projection matrix into the nullspace of h21.
Thus, Equation (8) reduces to projecting the received signal into the nullspace of h21 (i.e. to the space orthogonal to the interference vector) followed by applying zero-forcing (ZF) in the signal space that is orthogonal to h21. If H11P1 corresponds to single layer transmission, ZF is equivalent to spatial matched filtering which in this case is MRC with respect to the component of desired signal vector H11P1 in the space orthogonal to h21.
Post projection operation, the received signal vector can be written asG1y1=x1+(P1*H11*Ph21⊥H11P1)−1P1*H11*Ph21⊥n1.  (9)Therefore, the post-MMSE SINR for layer l (Equation (9)) can be written asSINR1,k=1/[(P1*H11*Ph21⊥H11P1)−1]l,l  (10)
The maximum number of layers that can be transmitted to UE1 from TP1 isνmax=rank(H11*Ph21⊥H11).  (11)
The precoding matrix that maximizes mutual information with a MMSE receiver has the form P1=U(:, 1:νmax), U is the unitary matrix whose first n columns correspond to the eigenvectors of H11*Ph21⊥H11 corresponding to the largest n eigenvalues.
With finite alphabet codebook, RI/PMI/CQI feedback can be carried out to maximize either the sum rate (over different layers) or the predicted total throughput                1. For a hypothesis on rank (i.e. number of layers, ν=1, . . . , rank (P1*H11*Ph21⊥H11P1)), the UE (e.g., UE1) finds the best precoding matrix (or vector for ν=1), denoted as P1opt, ν.        2. UE determines the information-theoretic mutual information (or sum rate across all layers) for ν layers as        
                    sum        ⁢                  -                ⁢                                  ⁢                  rate          v                    =                        ∑                      l            =            1                    v                ⁢                                  ⁢                              log            2                    ⁡                      [                          1              /                                                [                                                            (                                                                        P                          1                          *                                                ⁢                                                  H                          11                          *                                                ⁢                                                  P                                                      h                            21                                                    ⊥                                                ⁢                                                  H                          11                                                ⁢                                                  P                          1                                                                    )                                                              -                      1                                                        ]                                                  l                  ,                  l                                                      ]                                ,                  ⁢    where              P      1        =                  P        1                  opt          ,          v                    .      UE1 recommends the rank ν* that maximizes the sum rate sum-rateν as the preferred RI and P1opt,ν* as the preferred PMI.
In an embodiment of the invention, in which the SNR is finite, for non-negligible residual co-channel interference, the MMSE operation (Equation 3) may not be able to be decomposed into a projection operation followed by ZF. For a finite SNR (i.e. non-negligible residual co-channel interference), projecting the received vector into the nullspace of h21 may cause performance loss. Mean-square error (MSE) matrix is:E=E({circumflex over (x)}1−x1)({circumflex over (x)}1−x1)*=(I+P1*P1*R1−1P1)−1  (12)Defining:αl=[(I+P1*R1−1P1)−1]l,l,  (13)
The SINR associated with l-th layer after MMSE filtering is given by SINR1,l=1/αl−1. It can be shown that the precoder that maximizes the information-theoretic mutual information (or sum-rate across all layers)
      sum    ⁢          -        ⁢                  ⁢    rate    =            ∑              l        =        1            L        ⁢                  ⁢                  log        2            ⁡              [                  1          +                      SINR                          1              ,              l                                      ]            has the form P1=U1[:, 1:ν1]Σ1, where U1 is the unitary matrix whose first n columns are the eigenvectors of H11*R1−1H11 corresponding to the n largest eigenvalues and ν1 is the transmission rank chosen by TP1 for UE1 and Σ1 is a ν1×ν1 diagonal matrix.
The term “mutual information” used above may be conceptualized as follows: mutual information (R) is the maximum rate at which reliable communication (defined as the ability to transmit and receive k-bits of information using n-bits of coded bits so that the probability of bit error, Pe of receiving the b-bits incorrectly, tends to zero while the ratio k/n tends to a fixed limit that is less than or equal to the mutual information R. In other words, Pe→0 and k/n→r<=R) is possible. Conversely, there exists an encoding scheme which allows transmission of b-bits of information using n-bits of encoded data with vanishing error rate as long as k/n<=R.
Since R1=σ12I+H21P2P2*H21*, a method is to choose P2=U2[:, 1:ν2]|Σ2 where U2 is the unitary matrix whose first n columns are the right singular vectors of H21 corresponding to the n smallest singular values (including zero singular values) and ν2 is the transmission rank chosen by TP2 for UE2 and Σ2 is a ν2×ν2 diagonal matrix.
For codebook feedback (e.g., as in LTE Rel-10), unconstrained unitary matrices U1 and U2 can be determined by UE1.
The actions that a UE may take to mitigate downlink interference in one embodiment of the invention may be summarized as follows:
Step 1:
UE1 can find the “closest” codebook precoding matrix of rank ν1 by minimizing the subspace distance metric
  min            P      j        ∈          C      ⁡              (                  v          1                )            dj where
                              d          j                =                              1            2                    ⁢                                                                                                        P                    j                                    ⁢                                      P                    j                    *                                                  -                                                                            U                      1                                        ⁡                                          [                                                                        :                                                ,                                                  1                          ⁢                                                      :                                                    ⁢                                                      v                            1                                                                                              ]                                                        ⁢                                                            U                      1                      *                                        ⁡                                          [                                                                        :                                                ,                                                  1                          ⁢                                                      :                                                    ⁢                                                      v                            1                                                                                              ]                                                                                                          F            2                                              (        13        )            and C(ν1) is the codebook of rank v1. For a Rank 1 codebook, the Euclidean distance metric can be instead of the subspace distance metric. This constitutes feedback corresponding to the P1 matrix.Step 2:In a similar manner, UE1 can determine closest codebook precoding matrix of rank ν2 by minimizing the subspace distance metric arg
  min            P      j        ∈          C      ⁡              (                  v          2                )            dj where
      d    j    ==            1      2        ⁢                                                              P              j                        ⁢                          P              j              *                                -                                                    U                2                            ⁡                              [                                  :                                      ,                                          1                      :                                              v                        2                                                                                            ]                                      ⁢                                          U                2                *                            ⁡                              [                                  :                                      ,                                          1                      :                                              v                        2                                                                                            ]                                                                F      2      and C(ν2) is the codebook of rank ν2. This constitutes feedback corresponding to the P2 matrix.
In another embodiment of the invention, instead of determining an unconstrained U1, UE1 can determine P1εC(ν1) by selecting the precoding matrix that maximizes the sum-rate over all ν1 layers for a chosen P2 Suppose that the criterion arg
  min            P      j        ∈          C      ⁡              (                  v          2                )            dj in step 2 above selects the precoding matrix P2εC(ν2) UE1 can determine the sum rate
            sum      ⁢              -            ⁢                          ⁢      rate        =                  ∑                  l          =          1                          v          1                    ⁢                          ⁢                        log          2                ⁡                  [                      1            +                          SINR                              1                ,                l                                              ]                      ,under the assumption that TP2 is employing the precoding matrix P2, for each P1εC(ν1). UE1 can select P1 that maximizes the sum rate. This maximization can be carried out over different assumptions on the rank ν1=1, 2, . . . , N.
According to an embodiment of the invention, UE1 can report the pair (P1(1), P2(1)) to the TP. Thus, instead of TP2 muting its transmissions on some PRBs dynamically to eliminate complexity like for example in DPS/DPB, TP2 can ensure that the interference as seen by UE1 due to its transmission is quasi-orthogonal to the signal subspace of the TP1→UE1 link by signaling on the column space of P2(1). It is possible that TP2 cannot effectively find a user UE2 to transmit to if UE1 recommends just one signaling matrix P2(1). However, the probability that a suitable UE2 can be found may be increased if UE1 recommends more than one pair (P1(1), P2(1)). For example, UE1 can send a second pair ((P1(2), P2(2))), where P1(2) and P2(2) are respectively the second closest precoding matrices to U1[:, 1: ν1] and U2[:, 1:ν2] with respect to the subspace distance metric defined earlier. This allows TP1 and TP2 to jointly determine the precoding matrices to maximize the sum rate and also take into account other needs such as QoS, fairness, etc.
An example of how an embodiment of the invention is implemented will now be described. In this example, there is 4×2 antenna deployment where TP1 and TP2 each have 4 Tx antennas and UE1 and UE2 each have 2 Rx antennas. In this example, there is a rank restriction on the transmission to UE2 ν2=1. The TP2 to UE1 link is attenuated by 3 dB relative to TP1 to UE1 link. Similarly, the TP1 to UE2 link is attenuated by 6 dB relative to TP2 to UE2 link. In other words, both UE1 and UE2 are cell-edge users with respect to TP1 and TP2. UE1 and UE2 use the LTE Rel-10 4×1 and 4×2 codebooks which are based on Householder matrices to select PMI.
In this example, UE1 recommends rank 1 (P1(1), P2(1)) matrices based on the LTE Rel-10 4×1 codebook. The TP1 signals on P1(1) and TP2 signals on P2(1). The sum rate for UE1 and UE2 is computed as the sum of per-UE mutual information (that is, the mutual information (for UE1 and UE2) divided by 2) assuming MMSE receivers at UE1 and UE2.
In an embodiment of the invention, channel estimation may be used to determine the elements of the channel response matrix H. For example, TP1 and TP2 may transmit a first reference signal (TP1) and a second reference signal (TP2), which the UE1 receives. The first and second reference signals are transmitted according to a periodic schedule. The schedule, which comprises the periodicity and frame, subframe, or timeslot offset of the transmission occasions of the reference signals with respect to a certain time reference (e.g., system frame number, SFN=0) is signaled to the UE semi-statically (through radio resource control (RRC) messages or system broadcast) or dynamically (through downlink control information (DCI) within the physical downlink control channel (PDCCH)). The first and the second reference signal can be one or more of a CSI-RS, a CRS, a DMRS, a dedicated reference signal (DRS) or any other reference signal. Each reference signal is characterized by a time-frequency reuse pattern that indicates which set of subcarriers within each physical resource block (PRB) of a subframe is used for transmitting the reference signal. For example, CSI-RS is transmitted from 1, 2, 4 or 8 antenna ports and each port occupies 1 out of 12 subcarriers in 1 OFDM symbol of a subframe configured for CSI-RS transmission. The serving TP (e.g., TP1) can transmit CSI-RS resource information corresponding to both TP1 and TP2 and additionally other relevant TPs.
Although, for simplicity, the above example is phrased as if a first CSI-RS maps to TP1 and a second CSI-RS maps to TP2, it is understood that a CSI-RS may originate from multiple TPs, but show up in a DL subframe as a single CSI-RS resource. As far as the UE is concerned, the exact source or sources of the CSI-RS is transparent, and may be detected as a single CSI-RS being sent over a single CSI-RS resource.
Continuing with the example, UE1 can use the CSI-RS resource configuration and receive CSI-RS transmissions from at least TP1 and TP2. Based on the received CSI-RS transmissions, UE1 can determine the individual channel matrices corresponding to the radio links between TP1→UE1 and TP2→UE1. In one embodiment, UE1 uses MMSE channel estimation approach (e.g., as in the channel estimation described above) to determine the individual entries of the channel matrix H11, which corresponds to TP1→UE1 downlink channel and also the individual entries of the channel matrix H12, which corresponds to TP2→UE1 downlink channel. In another embodiment, UE1 uses DFT-based method or least squares (LS) method to determine H11 and H12.
In another embodiment of the invention, instead of selecting a second precoding matrix P2 of the pair (P1,P2), UE1 can simply select P1 after averaging the interference covariance matrix over different possible realizations of P2.
As noted above, the MMSE filter at UE1 can be written asG1=(I+P1*H11*R1−1H11P1)−1P1*H11*R1−1,whereR1=σ12I+H21P2P2*H21*.
Instead of computing R1 based on the selected value of P2, UE1 computes an expectation R1=σ12I+E[H21P2P2*H21*] where the expectation is over different possible realizations of P2. The expectation can be computed as an in time (subframes) and/or frequency (subbands).
Continuing with another embodiment of the invention, isotropic interference emulation is generally based on assuming spatially white transmissions from the interference node(s). However, such an assumption may not be optimal for precoder selection when interference rank is less than full rank. The TP can, alternatively, query the UE1 to report PMI/RI/CQI under different interference ranks as described below.                a. If the received signal is y1=H11P1x1+H21P2x2+n1, the interference covariance matrix is given by R1=σ12I+E[H21P2P2*H21*]. assuming i.i.d. input at TP2.        b. Isotropic interference emulation (prior art) models spatially white transmission from TP2, i.e. P2P2H=I.In general, P2εC, where        
  C  =            ⋃              v        =        1            N        ⁢          C              (        v        )            and C(ν) is the codebook of rank ν. So, the interference covariance can be computed as
            R      1        =                            σ          1          2                ⁢        I            +                        (                                    1              /              N                        ⁢                                        C                                              )                ⁢                              ∑                          l              =              1                                                    C                                              ⁢                                          ⁢                      E            ⁡                          [                                                H                  21                                ⁢                                  P                  l                                ⁢                                  P                  l                  H                                ⁢                                  H                  21                  H                                            ]                                            ,where P1εC.E[H21PlPlHH21H] can be computed by averaging the channel realization over the subband of interest. In other words, if H21,j is the channel response matrix in the j-th subcarrier of the subband and there are Jsc=12JPRB subcarriers in the subband comprising JPRB PRBs,
      E    ⁡          [                        H          21                ⁢                  P          l                ⁢                  P          l          H                ⁢                  H          21          H                    ]        ≈            (              1        /                  J          sc                    )        ⁢                  ∑                  j          =          1                          J          sc                    ⁢                          ⁢                        H                      21            ,            j                          ⁢                  P          l                ⁢                  P          l          H                ⁢                  H                      21            ,            j                    H                    leading to
      R    1    ≈                    σ        1        2            ⁢      I        +                  (                              1            /            N                    ⁢                                          ⁢                      J            sc                    ⁢                                  C                                      )            ⁢                        ∑                      l            =            1                                            C                                      ⁢                                  ⁢                              ∑                          j              =              1                                      J              sc                                ⁢                                    H                              21                ,                j                                      ⁢                          P              l                        ⁢                          P              l              H                        ⁢                                          H                                  21                  ,                  j                                H                            .                                          
Given that full rank transmission (i.e., rank=N) is likely not used by the interfering nodes all the time specially at the cell edge, instead of computing the interference covariance under isotropic assumption, UE1 can assume a restricted rank interference and report the CQI/PMI/RI leading to an improved interference emulation. The TP can semi-statically configure the rank or set of ranks for TP2 the UE should assume while computing the interference covariance matrix
      R    1    =                    σ        1        2            ⁢      I        +                  (                              1            /            N                    ⁢                                          ⁢                                  C                                      )            ⁢                        ∑                      l            =            1                                            C                                      ⁢                                  ⁢                              E            ⁡                          [                                                H                  21                                ⁢                                  P                  l                                ⁢                                  P                  l                  H                                ⁢                                  H                  21                  H                                            ]                                .                    The TP can also configure a codebook subset restriction that specifies which subset of C the UE shall use for computing the interference covariance matrix.
Referring back to FIG. 1 and to the flowchart of FIG. 6, an example of how a UE and a communication network may interact according to an embodiment of the invention will now be described. At step 601, TP1 transmits a CSI-RS configuration for TP2 and a CSI-RS configuration for TP2 to UE1. At step 602, UE1 receives a codebook subset restriction for CSI-RS configuration for TP2 and a codebook subset restriction for CSI-RS configuration for TP3. At step 603, UE1 receives a rank restriction for TP2 and a rank restriction for TP3. At step 604, UE1 receives a first CSI-RS from TP2 and a second CSI-RS from TP3. At step 605, UE1 determines a (P1, P2) pair subject to the codebook subset restriction and rank restriction. At step 606, UE1 sends a CSI report back to TP1 which includes information pertaining to (P1, P2). Again, the first or second CSI-RS may be transmitted by a combination of multiple TPs, but will be perceived by UE1 as a first CSI-RS resource and a second CSI-RS resource (i.e., as a first CSI-RS and a second CSI-RS).
It can be seen from the foregoing that a novel method and apparatus for mitigating downlink control information has been described. It is to be noted that embodiments within the scope of the present disclosure may also include computer-readable media for carrying or having computer-executable instructions or data structures stored thereon. Such computer-readable media can be any available media that can be accessed by a general purpose or special purpose computer. By way of example, and not limitation, such computer-readable media can comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to carry or store desired program code means in the form of computer-executable instructions or data structures. When information is transferred or provided over a network or another communications connection (either hardwired, wireless, or combination thereof) to a computer, the computer properly views the connection as a computer-readable medium. Thus, any such connection is properly termed a computer-readable medium. Combinations of the above should also be included within the scope of the computer-readable media.
Embodiments may also be practiced in distributed computing environments where tasks are performed by local and remote processing devices that are linked (either by hardwired links, wireless links, or by a combination thereof) through a communications network.
Computer-executable instructions include, for example, instructions and data which cause a general purpose computer, special purpose computer, or special purpose processing device to perform a certain function or group of functions. Computer-executable instructions also include program modules that are executed by computers in stand-alone or network environments. Generally, program modules include routines, programs, objects, components, and data structures, etc. that perform particular tasks or implement particular abstract data types. Computer-executable instructions, associated data structures, and program modules represent examples of the program code means for executing steps of the methods disclosed herein. The particular sequence of such executable instructions or associated data structures represents examples of corresponding acts for implementing the functions described in such steps.
While the present disclosure and the best modes thereof have been described in a manner establishing possession by the inventors and enabling those of ordinary skill to make and use the same, it will be understood that there are equivalents to the exemplary embodiments disclosed herein and that modifications and variations may be made thereto without departing from the scope and spirit of the disclosure, which are to be limited not by the exemplary embodiments but by the appended claims.